JP2005331476A - Method and device for detecting three-dimensional information - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method and a device for detecting three-dimensional information, capable of efficiently reducing the noise in an image, serving as the basis for calculation of a distance image, within a prescribed time. <P>SOLUTION: A subject is irradiated switchingly with the first intensity modulated light, and the second intensity modulated light of which the ratio of intensity change to a time is different from that of the first intensity modulated light; the reflected light from the subject is photographed through a high-speed shutter to acquire the first image at the time, when irradiated with the first intensity modulated light; and the second image at the time when irradiated with the second intensity modulated light, the respective image accumulation numbers of the first image and the second image in an image accumulation type filter are controlled to reduce the noise, based respectively on an image signal level of the first image and an image signal level of the second image; and the distance image is calculated from an intensity ratio of the first image and the second image output from the image accumulation type filter. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a three-dimensional information detection method and apparatus, and more particularly to a three-dimensional information detection method and apparatus for detecting a depth shape of a subject and acquiring three-dimensional information.

  There are many methods for detecting a three-dimensional position or depth distance of an object (hereinafter, “depth distance” is simply referred to as “distance”), such as a triangulation method, an interference method, an optical flight time measurement method, and a moire method. However, there are few implementation examples of the technology for acquiring the distance image of the subject at high speed. For example, there are a method using a stereo camera (see Non-Patent Document 1), a light cutting measurement method (see Non-Patent Document 2), and the like.

Patent Document 1 describes that the distance is calculated at high speed by a method using intensity-modulated light and a high-speed shutter based on the measurement of time of flight of light.
JP 2000-121339 A S. Kimura, H .; Kano, T .; Kanade, A .; Yoshida, E .; Kawamura, and K.K. Oda, “CMU video-rate stereo machine,” in Proceedings of 1995 9-18 S. Yoshimura, T .; Sugiyama, K .; Yonmoto, and K.K. Ueda, “A 48 kframe / s CMOS image sensor for real-time 3-D sensing and motion detection,” in 2001 International Solid-State Circuits ISSCC Digg. of Tech. Papers (Institute of Electrical and Electronics Engineers, San Francisco, 2001), pp. 94-95,436.

  In order to detect a three-dimensional position from a method using a stereo camera, that is, from an image obtained by a stereo camera, complicated image calculation processing is required, so that it is difficult to increase the definition of the image. In addition to the problem of occlusion (viewpoint obstruction), there is a problem in that it is difficult to match images on a subject without a feature point, and distance cannot be detected.

  In the light cutting measurement method, that is, a method in which a slit-shaped laser beam is scanned, received by a CMOS sensor, and a distance is obtained according to the principle of triangulation, the slit-shaped laser beam is scanned and scanned. Department is necessary. Furthermore, in order to obtain the amount of reflected light necessary for the sensitivity of the sensor, it is generally necessary to irradiate a laser beam with a high light intensity. When the subject is a person, there is a problem in safety. Furthermore, since the camera and the light source are separated from each other, there is a problem that the distance of the subject portion that becomes the shadow of the illumination cannot be detected.

  On the other hand, the method using the intensity-modulated light and the high-speed shutter makes it easy to calculate the distance and does not require a mechanical scanning mechanism for the irradiation light. Furthermore, in this method, since intensity modulated light that increases / decreases relatively slowly is used instead of steep pulsed light, the frequency characteristics of intensity modulation are low compared to a laser light source, but an incoherent LED close to natural light ( (Light-emitting diode) Light can be applied, and safe shooting is possible even when the subject is a person.

  However, Patent Document 1 has a problem that a specific method for reducing noise in an image that is a basis for calculating a distance image is not considered.

  The present invention has been made in view of the above points, and an object of the present invention is to provide a three-dimensional information detection method and apparatus that can efficiently reduce noise of an image that is a basis for calculation of a distance image within a predetermined time. .

According to the first aspect of the present invention, the first intensity-modulated light and the second intensity-modulated light whose rate of intensity change with respect to time is different from the first intensity-modulated light are switched to irradiate the subject,
The reflected light from the subject is photographed through a high-speed shutter, and a first image at the time of irradiation of the first intensity-modulated light and a second image at the time of irradiation of the second intensity-modulated light are acquired,
Based on each of the video signal level of the first image and the video signal level of the second image, noise reduction is performed by controlling the image accumulation number of each of the first image and the second image in the image accumulation type filter,
By calculating the distance image from the intensity ratio between the first image and the second image output from the image storage filter, the noise of the image can be efficiently reduced within a predetermined time.

According to the second aspect of the present invention, the first intensity-modulated light and the intensity-modulated light irradiation for irradiating the subject by switching between the second intensity-modulated light whose intensity change rate with respect to time is different from that of the first intensity-modulated light. Means,
A photographing unit that includes a high-speed shutter and photographs reflected light from the subject to obtain a first image when the first intensity-modulated light is irradiated and a second image when the second intensity-modulated light is irradiated; ,
Recursive filter means for reducing noise by controlling the number of images stored in the first image and the second image based on the video signal level of the first image and the video signal level of the second image, respectively.
By having a distance image calculation means for calculating a distance image from the intensity ratio between the first image and the second image output from the recursive filter means, the invention of claim 1 can be realized, and the noise of the image is reduced within a predetermined time. It can be reduced efficiently.

The invention according to claim 3 is the three-dimensional information detection apparatus according to claim 2,
The recursive filter means can reduce image noise more efficiently by changing the filter coefficient according to the number of accumulated images of the first image and the second image.

The invention according to claim 4 is the three-dimensional information detection apparatus according to claim 2 or 3,
The intensity-modulated light irradiating means switches the first intensity-modulated light and the second intensity-modulated light based on the number of accumulated images of the first image and the second image and irradiates the subject with the first intensity-modulated light irradiation means. The number of images stored in the image and the second image can be varied.

  According to the present invention, it is possible to efficiently reduce noise in an image that is a basis for calculating a distance image within a predetermined time.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 shows a block diagram of an embodiment of a three-dimensional information detection apparatus to which the method of the present invention is applied. In the figure, the signal generator 10 generates a light intensity modulation signal such as a triangular wave and supplies it to the switch 12. The switch 12 has two output terminals, one output terminal is directly connected to the light source unit 16, and the other output terminal is connected to the light source unit 16 via the delay line 14. By the switching of the switch 12, light intensity modulation signals whose modulation phases are different by 180 degrees, for example, are supplied to the light source unit 16.

  The light source unit 16 irradiates the subject 18 with intensity modulated light 17 modulated in accordance with the light intensity modulation signal, and the reflected light 19 from the subject 18 is photographed by the camera 20 with a high-speed shutter. The video signal output from the camera 20 is switched by the switch 22 and supplied to one of the filter processing units 24 and 26. The switch 22 is synchronized with the switch 12, and the images at the time when the intensity-modulated light different in phase by 180 degrees is irradiated are supplied to the filter processing units 24 and 26, respectively.

  Each of the filter processing units 24 and 26 performs a filter process on each image at the time of irradiation with intensity-modulated light having a phase difference of 180 degrees, and supplies each image after the filter process to the distance image calculation device 28 and after the filter process. A switching control signal is generated from the brightness value of each image and supplied to the switches 12 and 22. The distance image calculation device 28 obtains a distance image from each image after filtering (an image at the time of irradiation of intensity-modulated light having a phase difference of 180 degrees), and outputs it as a video signal of the distance image.

  Here, since the main noise factor of the video signal output from the camera 20 is shot noise of the image intensifier used for the high-speed shutter, noise can be reduced by performing image accumulation. In the image accumulation addition, the noise component is reduced to 1 / (√N) times the number N of accumulated images. However, if the number of images is large in the simple image accumulation and addition method, the memory is saturated.

  Therefore, when a subject that moves to some extent is targeted, a recursive filter that is one of image accumulation filters (Reference 1: RH Mcmann, S. Kreinik, JK Moore, A. Kaiser, and J Rossi, “A digital noise reducer for encoded NTSC signal” J. SMPTE, Vol. 87, pp. 129-133, March 1978) is effective.

  This recursive filter does not saturate the memory even when the number of images increases, and the detection time constant can be varied by the filter coefficient, so it can be used for noise reduction in moving image shooting (Reference 2: Yuichi Ninomiya, Yoshimichi Otsuka “ Motion correction type noise reducer, “Journal of Television Society, Vo 1.39, No. 10, pp. 956-962, October 1985).

In the present invention, each of the filter processing units 24 and 26 is configured by a recursive filter, and the filter processing unit 24 filters the first image V ₊ captured when the intensity of the intensity-modulated light increases with time. Then, the filter processing unit 26 filters the second image V ₋ taken at the time of the decreased modulated light in which the light intensity of the intensity modulated light decreases with time.

  FIG. 2 shows a block diagram of an embodiment of the filter processing units 24, 26. Each of the filter processing units 24 and 26 is a subtracter 30 that subtracts the video signal of the image output from the frame memory 36 from the video signal of the captured image of the camera 20, and the inverse of the filter coefficient n to the output signal of the subtractor 30. A multiplier 32 for multiplying the coefficient 1 / n, an adder 34 for adding the video signal of the image output from the frame memory 36 to the output signal of the multiplier 32, and a frame memory 36 for storing the output signal of the adder 34 , And a distributor 38 for distributing the output signal of the adder 34 to the frame memory 36 and the output to the outside.

Assuming that the video signal output from the camera is V ₊ , the output signal of the filter processing unit 24 is expressed by equation (1).

ここで、Ｆ_ｔ−１はフレームメモリ３６の出力信号、１／ｎは乗算器３２で乗算される係数である。このフィルタ処理をｔ回繰り返すことで、ノイズ成分δは、（２）式で表わされるδ’に低減される。

Here, F _t−1 is an output signal of the frame memory 36, and 1 / n is a coefficient multiplied by the multiplier 32. By repeating this filtering process t times, the noise component δ is reduced to δ ′ expressed by equation (2).

（２）式からｔが十分に大きな場合、ノイズ成分は、（２ｎ−１）^−１／２倍に低減される。

When t is sufficiently large from the equation (2), the noise component is reduced to (2n-1) ^-1/2 times.

On the other hand, in the present invention, the captured image V ₊ = V ₁ ± δ ₁ in the case of increasing modulated light and the captured image V ₋ = V ₂ ± δ _{2 in the} case of decreasing modulated light (V ₁ and V ₂ are images that do not contain noise. , Δ ₁ and δ ₂ are noise components), the distance image D is expressed by the equation (3).

ここで、撮影画像Ｖ_＋，Ｖ₋を、フィルタ処理部２４，２６を透過させた後の距離画像ノイズ成分δ_Ｄ’は、（４）式で表わされる。

Here, the distance image noise component δ _D ′ after the captured images V ₊ , V ₋ are transmitted through the filter processing units 24, 26 is expressed by equation (4).

なお、ここで、ｔ_１，ｔ_２は、各画像Ｖ_＋，Ｖ₋の画像蓄積数、つまり、フィルタ処理部２４，２６それぞれのフィルタ処理回数である。

Note _that, t 1, _{t 2,} each image _V +, V _- image storage number, that is, the filter processing unit 24, 26 is a respective filter processing times.

In general, since the image V ₊ and the image V ₋ have different SN ratios, it is effective to intensively improve an image having a low SN ratio in order to efficiently reduce noise within a limited time. For this purpose, the image accumulation numbers t ₁ and t ₂ are optimally controlled with reference to the luminance value or SN ratio of each image.

Next, the effect when the number of acquired images is optimized is estimated as follows. Image _V +, V _- if the ratio of the SN ratio is _{δ 2 / V 2 = k (} δ 1 / V 1), the noise reduction factor r before and after the filter processing is represented by equation (7).

画像Ｖ_＋，Ｖ₋の画像蓄積数ｔ_１，ｔ_２が同じ場合（ｔ_１＝ｔ_２＝Ｔ／２、ただしＴは画像の総蓄積数）、ａ_１＝ａ_２であるため、ノイズ低減率ｒは（８）式で表わされる。

When the image accumulation numbers t ₁ and t _{2 of the} images V ₊ and V ₋ are the same (t ₁ = t ₂ = T / 2, where T is the total accumulation number of images), since a ₁ = a ₂ , noise reduction The rate r is expressed by equation (8).

一方、画像Ｖ_＋，Ｖ₋の画像蓄積数ｔ_１，ｔ_２が異なる場合、ノイズ低減率ｒは（９）式で表わされる。

On the other hand, the image _V +, V _- the image storage number _t 1 of, _{if t 2} are different, the noise reduction factor r is expressed by equation (9).

このとき、二乗ノイズ低減率ｒ^２を最小とする画像蓄積数ｔ_１の値は（１０）式から得られる。また、画像蓄積数ｔ_２はｔ_２＝Ｔ−ｔ_１から得られる。

At this time, the value of the image accumulation number t ₁ that minimizes the square noise reduction rate r ² is obtained from the equation (10). The image storage number _{t 2} is obtained from _t 2 = _{T-t 1.}

一例として、ＳＮ比の比率ｋ＝３、係数１／ｎ＝１／３２、画像の総蓄積数Ｔ＝５０の場合、画像蓄積数ｔ_１を最適化しないときはノイズ低減率ｒ＝０．４８であるのに対し、（１０）式で最適化したときはノイズ低減率ｒ＝０．３８と改善される。このときの画像蓄積数ｔ_１＝８、画像蓄積数ｔ_２＝４２である。

As an example, in the case of the SN ratio k = 3, the coefficient 1 / n = 1/32, and the total image accumulation number T = 50, the noise reduction rate r = 0.48 when the image accumulation number t ₁ is not optimized. On the other hand, when the optimization is performed by the expression (10), the noise reduction rate r = 0.38. At this time, the image accumulation number t ₁ = 8 and the image accumulation number t ₂ = 42.

If the shooting conditions of the camera 20 are constant, the luminance signal level and noise level of the image shot by the camera 20 have a one-to-one relationship. In other words, the luminance signal level of the image has a one-to-one relationship with the SN ratio. FIG. 3 shows an example of measurement of the luminance level versus SN ratio of an image. If this relationship is measured in advance, the SN ratio and SN ratio k of the captured image can be obtained from the luminance signal level of the captured image, and the image accumulation numbers t ₁ and t ₂ can be obtained from the equation (10). .

Therefore, the image accumulation numbers t ₁ and t ₂ are determined from the SN ratio of the images V ₊ and V ₋ acquired first. The number of image accumulations t _{1 of} each photographed image V ₊ , V _{− based} on the luminance level of the image F ₁ at the time of increasing modulated light and decreasing modulated light, which are first acquired and stored in the frame memories 36 of the filter processing units 24, 26. , T ₂ can be determined. Note that the value of the subject portion of interest in the image is used as the luminance signal level at this time.

As shown in FIG. 1, the image accumulation number of each image is controlled by obtaining the optimum image accumulation numbers t ₁ and t ₂ based on the output values of the filter processing units 24 and 26, respectively. Supply control signals. The switch 12 switches the intensity modulation signal in accordance with the control signal, and in synchronization with this, the switch 22 switches the captured images V ₊ and V ₋ of the camera 20 in accordance with the control signal.

In the above description, + factor 1 / n is photographed image _V, V _- was described as the same for the coefficient _{1 /} n 1 in accordance with this factor 1 / n in the image storage number _t 1, _{t 2} respectively, 1 / By changing to n ₂ , the maximum noise reduction can be performed within the image accumulation numbers t ₁ and t ₂ .

When the image accumulation number t ₁ is given in the equation (5), the value of the coefficient 1 / n ₁ that minimizes a ₁ is determined. The relationship between the image accumulation number t ₁ and the filter coefficient n ₁ is as shown in FIG. The relationship shown in FIG. 4 can be approximated by n ₁ = 0.32t ₁ +2.55 in the region where the image accumulation number t ₁ is 100 or less. The same applies to the equation (6). When the image accumulation number t ₂ is given, the value of the coefficient 1 / n ₂ that minimizes a ₂ is determined.

When the effect is calculated, when the S / N ratio k = 3, the coefficient 1 / n = 1/32, the total image accumulation number T = 50, the image accumulation number t ₁ = 8, and the image accumulation number t ₂ = 42, If the value of the coefficient 1 / n is not changed, the noise reduction rate r = 0.38. On the other hand, when the coefficient 1 / n ₁ = 5.1 and the coefficient 1 / n ₂ = 16.0 are optimized, the noise reduction rate r = 0.20, and the coefficient 1 / n is optimized. I understand that.

  FIG. 5 shows a specific block diagram of an embodiment of a three-dimensional information detection apparatus to which the method of the present invention is applied. In the figure, an LED array light source having an output light intensity of 1 W and a wavelength of 850 nm was used as the light source unit 40, and the subject was irradiated with intensity modulated light 41 that was intensity modulated at a frequency of 45 MHz. The reflected light 43 is collected by the camera lens 44 and imaged and input to the image intensifier 48 by the dichroic prism 46.

  The image intensifier 48 has a high-speed shutter mechanism only for near-infrared light, and performed a shutter operation with a time width of 2 nsec and a repetition frequency of 45 MHz. The output light from the fluorescent screen of the image intensifier 48 is photographed by the distance detection CCD camera 50, supplied to the personal computer 52 for data acquisition, and the software on the personal computer 52 uses the filter processing units 24, 26. Filter processing was performed.

On the other hand, the visible reflected light 43 passes through the dichroic prism 46 and is photographed by the color camera 54 to output a normal color image video signal. The object 42 used was a flat plate 2 m ahead, and captured images V ₊ and V ₋ output from the CCD camera 50 were captured with a sampling resolution of 10 bits. Photographed image _V +, V _- each of the luminance signal level 392mV, 205mV, photographed image _V +, V _- each of the noise level was conducted 6.3MVrms, the measurement in the case of 9.5MVrms.

FIG. 6 shows the relationship between the total accumulation number T (= t ₁ + t ₂ ) of the images V ₊ and V ₋ and the noise reduction rate r of the distance image. Here, the image storage number _t 1, _{t 2} and the filter coefficients _n 1, _{n 2} together optimized if not the calculation result (dotted line) and experimental results (black square points), image storage number _t 1, _{t 2} only optimization calculations (dashed line) in the case of and experimental results (solid triangles point), the image storage number t _1, t ₂ and the filter coefficients n _1, n ₂ together optimized if the calculation result (two-dot chain line) And the experimental results (black dots) are shown.

  As a result, when the total accumulated number T of images is 16 frames, the noise reduction rate r of the experimental result is 0.68 when not optimized, whereas the noise reduction of the experimental result is optimized by optimizing only the accumulated number of images. The rate r was reduced to 0.59, and the noise reduction rate r as an experimental result was reduced to 0.39 by optimizing the number of accumulated images and filter coefficients.

  Thus, according to this embodiment, noise can be efficiently reduced within a predetermined time. By reducing this noise, the distance detection resolution, which was conventionally several centimeters, can be improved to the order of several millimeters, and can be applied to a wide range of fields such as shape measurement.

  Here, an example of the distance image calculation process executed by the distance image calculation device 28 will be briefly described. Here, a description will be given of a distance detection method using light that increases and decreases with time as intensity-modulated light and a camera having a high-speed shutter mechanism (corresponding to the image intensifier 48 and the CCD camera 50). The image intensifier 48 is composed of a photoelectric conversion surface, a microchannel plate (MCP), and a phosphor screen, and controls the value of the applied voltage between the photoelectric conversion surface and the MCP, thereby increasing the imaging sensitivity with time. It is possible to change the applied voltage, and the gate voltage can be captured in a short time by making the applied voltage into a pulse shape.

  As shown in FIG. 7A, the subject placed at the distance d is irradiated with intensity-modulated light A1 whose light intensity increases with a factor s over time, and the reflected light A2 from the subject is pulsed at time ts. When the image is captured for a short time with the gain A3, the signal amount E + (d, ts) detected by the CCD camera 50 is expressed by equation (11).

ここで、ＴＬはカメラレンズなどのレンズ光学系の透過率、ρは被写体の表面の反射特性係数、Ｆ０は光の最大照射強度、Δｔは撮像時間幅であり、光変調周期に対して十分小さい値である。また、ｃは光速、２ｄ／ｃはＣＣＤカメラ５０から被写体までの距離ｄを光が往復する時間、ｌはカメラレンズから被写体までの距離であり、式の分母は光の拡散による減衰を考慮した項である。

Here, TL is the transmittance of a lens optical system such as a camera lens, ρ is the reflection characteristic coefficient of the surface of the object, F0 is the maximum irradiation intensity of light, and Δt is the imaging time width, which is sufficiently small with respect to the light modulation period. Value. Also, c is the speed of light, 2d / c is the time for light to travel back and forth between the distance d from the CCD camera 50 to the subject, l is the distance from the camera lens to the subject, and the denominator takes into account attenuation due to light diffusion. Term.

  Next, as shown in FIG. 7B, the intensity-modulated light A4 whose light intensity decreases with a coefficient s with time is irradiated, and the reflected light A5 from the subject is imaged for a short time with the pulsed imaging gain A3 at time ts. In this case, the signal amount E− (d, ts) detected by the CCD camera 50 is expressed by equation (12).

ここで、Ｔ_Ａは光強度の変調周期である。なお、１回の撮像では感度が不十分である場合は、１フィールド内に、撮像ゲインの変調周波数と同等の繰り返しパルス光を照射し、蓄積型の撮像素子で蓄積し十分な感度を確保する。なお、図７（Ｂ）の強度変調光は図７（Ａ）の強度変調光と連続して送出しても良い。

Here, T _A is the modulation period of the light intensity. If the sensitivity is insufficient for one imaging, a pulsed light equivalent to the modulation frequency of the imaging gain is irradiated in one field and accumulated by a storage type imaging device to ensure sufficient sensitivity. . Note that the intensity-modulated light in FIG. 7B may be transmitted continuously with the intensity-modulated light in FIG.

  From equation (11) and equation (12), the intensity ratio R = E + / E− between two images having different light intensities is taken and the distance d is obtained as equation (13).

（１３）式で示されるように、光強度増加時と光強度減少時に撮像した２つの画像間の比Ｒを計算するだけで、被写体の反射率や光の拡散による光の減衰効果等の影響をキャンセルし、高速に距離を求めることができる。

As shown in the equation (13), only by calculating the ratio R between two images picked up when the light intensity is increased and when the light intensity is decreased, the influence of the reflectance of the subject, the light attenuation effect due to light diffusion, etc. Can be canceled and the distance can be obtained at high speed.

  The distance image calculation device 28 reads a subject image captured for a short time when the intensity-modulated light whose light intensity increases from the filter processing unit 24, and for a short time when the intensity-modulated light whose light intensity decreases from the filter processing unit 26. Read the captured subject image. The distance image calculation device 28 calculates the distance d by performing the calculation of the equation (13) after adjusting the luminance of the image of the read signal and the setup adjustment for adding a certain luminance level to the entire image. Then, a distance video signal expressing the distance in light and dark is obtained. Further, the distance image calculation device 28 converts the distance video signal into a desired television signal standard and outputs it.

  The signal generator 10, the switch 12, and the light source unit 16 correspond to the intensity-modulated light irradiation unit, the camera 20 with a high-speed shutter corresponds to the photographing unit, and the filter processing units 24 and 26 serve as the recursive filter unit. Correspondingly, the distance image calculation device 28 corresponds to the distance image calculation means.

  A three-dimensional information detection apparatus to which the method of the present invention is applied can be applied to various fields as a three-dimensional camera capable of acquiring depth distance information together with a color image of a subject. As an example, a television broadcast camera can be applied to extraction and synthesis of images based on distance information, and a stereoscopic television camera. In addition, three-dimensional modeling, medical diagnosis, observation in automobiles, aviation, space, marine engines and other equipment, inspection of gas pipes, exhaust outlets, water pipes, boilers, turbines, investigation of ancient tombs and ruins, Applicable to fields such as observation of animal ecology.

It is a block diagram of one Embodiment of the three-dimensional information detection apparatus to which this invention method is applied. It is a block diagram of one Embodiment of a filter process part. It is a figure which shows the example of a measurement of the luminance level versus SN ratio of an image. It is a figure which shows the relationship between the image accumulation number t and the coefficient 1 / n. It is a specific block diagram of one Embodiment of the three-dimensional information detection apparatus to which this invention method is applied. It is a figure which shows the relationship between the image accumulation number of an image, and the noise reduction rate of a distance image. It is a figure for demonstrating the distance detection method by the intensity | strength modulation light in which a pulse-shaped imaging gain and light intensity increase and decrease.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Signal generator 12,22 Switch 14 Delay line 16,40 Light source part 17 Intensity modulated light 18,42 Subject 19,43 Reflected light 20 Camera with high-speed shutter 24,26 Filter processing part 28 Distance image arithmetic unit 30 Subtractor 32 Multiplication Device 34 adder 36 frame memory 38 distributor 44 camera lens 46 dichroic prism 48 image intensifier 50 CCD camera 52 personal computer

Claims (4)

The first intensity-modulated light and the second intensity-modulated light whose rate of intensity change with respect to time is different from the first intensity-modulated light are switched to irradiate the subject,
The reflected light from the subject is photographed through a high-speed shutter, and a first image at the time of irradiation of the first intensity-modulated light and a second image at the time of irradiation of the second intensity-modulated light are acquired,
Based on each of the video signal level of the first image and the video signal level of the second image, noise reduction is performed by controlling the image accumulation number of each of the first image and the second image in the image accumulation type filter,
A three-dimensional information detection method, wherein a distance image is calculated from an intensity ratio between a first image and a second image output from the image storage filter. Intensity modulated light irradiating means for irradiating the subject by switching between the first intensity modulated light and a second intensity modulated light having a rate of intensity change with respect to time different from that of the first intensity modulated light;
A photographing unit that includes a high-speed shutter and photographs reflected light from the subject to obtain a first image when the first intensity-modulated light is irradiated and a second image when the second intensity-modulated light is irradiated; ,
Recursive filter means for reducing noise by controlling the number of images stored in the first image and the second image based on the video signal level of the first image and the video signal level of the second image, respectively.
3. A three-dimensional information detection apparatus comprising distance image calculation means for calculating a distance image from an intensity ratio between the first image and the second image output from the recursive filter means. The three-dimensional information detection apparatus according to claim 2,
3. The three-dimensional information detecting apparatus according to claim 1, wherein the recursive filter means varies a filter coefficient according to the number of accumulated images of the first image and the second image. The three-dimensional information detection apparatus according to claim 2 or 3,
The intensity-modulated light irradiating means switches the first intensity-modulated light and the second intensity-modulated light based on the number of accumulated images of the first image and the second image and irradiates the subject. Three-dimensional information detection device.

Images